The NI LabVIEW GPU Analysis Toolkit enables you to communicate with NVIDIA CUDA graphics processing units (GPUs) from LabVIEW applications. With this toolkit, you can open references to GPU devices in your system, transfer data between CPU and GPU memory, and control the execution of GPU code.
The LabVIEW GPU Analysis Toolkit features a wide variety of CUDA Basic Linear Algebra Subroutines library (cuBLAS) and CUDA Fast Fourier Transform library (cuFFT) signal processing functions wrapped in LabVIEW for quickly prototyping GPU algorithms. In addition, you can take advantage of device selection and resource management using CUDA Runtime and Driver APIs. For advanced users, the toolkit contains documentation on calling your custom GPU code from LabVIEW applications or executing functions from other freely available NVIDIA libraries such as the NVIDIA Performance Primitives library (NPP) and CUDA Sparse Matrix library (cuSPARSE).
You can use this toolkit with the LabVIEW 32-bit and 64-bit development environments on the Windows OS. For additional information on toolkit capabilities or to review the documentation, visit the Resources tab. Note that LabVIEW GPU Analysis Toolkit functions do not work with the LabVIEW Real-Time Module. You cannot use the toolkit to execute existing LabVIEW graphical code on GPU devices.
The primary delivery mechanism for the LabVIEW GPU Analysis Toolkit is via download; you can access the download for evaluation or after purchasing at ni.com/downloads.
For information on system requirements, view the Resources tab.